Scientists at Indiana University's Social Networking Observatory have shared research that has uncovered more than 1,000 AI spam bots designed to trick social network users and attempt to steal their profiles.
ChatGPT used to run botnets
According to research data, the bots were created using ChatGPT functions. Criminals take advantage of the AI capability to create texts that look trustworthy and run so-called "botnets" on Twitter/X. Networks of hundreds of malicious bots controlled in whole or in part by software, botnets run spam campaigns on social networks. And what is more worrying: they manage to go unnoticed by current anti-spam filters.
The research was carried out by computer science researcher Kai-Cheng Yang and computer science professor Filippo Menczer. According to the data collected, more than 1,000 bots active on X promote fraudulent cryptocurrencies and NFTs to convince users to invest or even steal their crypto wallets. To appear credible, the bots reply to messages with ChapGPT results and even post selfies stolen from other profiles.
In an interview with Insider, Menczer noted that the rise of social media has provided a cheap way for malicious actors to monetize fake content by reaching a large audience. And new AI tools exacerbate this situation. According to him, these functions "further reduce the cost of generating fake but trustworthy content on a large scale, defeating the already weak moderation defenses of social media platforms."
Whereas it was once easy to recognize some bots by their more robotic behavior, today bots use generative AI to their advantage to produce content very similar to that of human users. For Yang, "the advancement of AI tools will permanently distort the idea of information on the Internet," he told Insider.
AI has been used to spoof news sites
Another target for criminals who have used generative AI tools in their scams are news sites. Using tools similar to ChatGPT, they can create fake news pages and spam.
ANewsGuard, a private company that evaluates the reliability of news and information sites, has been conducting an audit process since April. To date, it has found more than 400 AI-generated news pages. In addition to being able to have a specific purpose for these fake news, these pages benefit from automated advertising revenue.
Regulators are racing against the clock to keep pace with this process and find viable and effective ways to detect AI-generated content. However, AI resources seem to keep running faster, releasing more news than the regulatory search can keep up with.
You can read:
Leave a Reply
You can also Read: